# On the parallelization of matrix multiplications in Mathematica 8

I have installed Mathematica 8, but I think the commands for parallelizations do not work! Even when I try to test the example in the Help of Mathematica, I face with

ParallelDo::nopar: No parallel kernels available; proceeding with sequential evaluation.

Please guide me how to resolve this problem. In fact I wish to then use ParalledDo in the following code to reduce the computational time:

ClearAll[n, A, V, Id, b, L, i, pr]
n = 1000;
A = SparseArray[{{i_, i_} -> 3 i, {i_, j_} /; i == j + 1 -> j}, {n,
n}];
b = SparseArray[Table[1, {i, n}]];
DA = Diagonal[A];
B = SparseArray[Table[(1/DA[[i]])^1, {i, 1, n}]];
Id = SparseArray[{{i_, i_} -> 1}, {n, n}];
V = SparseArray[DiagonalMatrix[B]];
ParallelDo[
VV = SparseArray[A.V];
V = SparseArray[V].SparseArray[2 Id - VV];
L[i] = Norm[N[SparseArray[b - SparseArray[A].SparseArray[V.b]]]];
Print["The residual norm of the linear system solution is:"
Column[{i}, Frame -> All, FrameStyle -> Directive[Blue]]
Column[{L[i]}, Frame -> All, FrameStyle -> Directive[Blue]]];
, {i, 4}] // Timing


I will be thankful if anyone gives me some tips to reduce the computational cost of matrix-by-matrix multiplications in the above written code.

-

## migrated from superuser.comMay 28 '12 at 10:24

This question came from our site for computer enthusiasts and power users.

BTW the nopar message suggests that you either have a single core CPU or you set Mathematica not to launch any parallel kernels. You can launch kernels manually using LaunchKernels even on a single core machine. – Szabolcs May 26 '12 at 19:45

You need to make sure Local Kernels are not disabled. If they're enabled and the options set as shown, then evaluating, for example, ParallelDo[Pause[1]; f[i], {i, 4}] without parallel kernels launched automatically launches them.